Inertial Stabilized Platforms (ISPs) are critical for maintaining the stability of sensors and cameras on dynamic vehicles like Unmanned Ground Vehicles (UGVs). These systems must counteract complex disturbances, including friction, mass imbalance, and vehicle vibrations. While Active Disturbance Rejection Control (ADRC) has emerged as a robust solution for handling these uncertainties, standard implementations face limitations due to theoretical gaps in disturbance compensation and sensor-induced delays. This paper, published in IEEE Access, proposes a modified ADRC scheme (ADRC-A) that addresses these "intrinsic" weaknesses.
The Problem: Observation Error and Sensor Lag
Standard ADRC (referred to as ADRC-C) relies on an Extended State Observer (ESO) to estimate and compensate for generalized disturbances. However, mathematical analysis reveals a residual error in steady-state performance under disturbance. Furthermore, rate gyros—the primary sensors for angular velocity—introduce internal dynamics or "lag." Standard ADRC treats these lagged outputs as the absolute truth, leading to delayed control actions and degraded precision.
Proposed Methodology: ADRC-A
The researchers introduced two key modifications to the first-order nonlinear ADRC to create the ADRC-A scheme:
- Corrected Disturbance Compensation: By analyzing the ESO's transfer function, the authors identified that the system output does not perfectly reach zero under disturbance. They corrected this by adding an error-related term ($\beta_1 E(s)$) to the compensation quantity, effectively driving the disturbance-induced error toward zero.
- Improved Speed Observer: To mitigate gyro lag, the team integrated an estimated gyro transfer function ($G_{sEst}(s)$) directly into the observer. This allows the controller to estimate the actual plant speed rather than relying on the delayed sensor signal.
To validate these changes, the team used the describing function method—a technique to linearize nonlinear elements for frequency-domain analysis—allowing them to compare the "dynamic stiffness" and stability margins of the old vs. new schemes.
Key Results and Experimental Validation
The modified ADRC-A scheme demonstrated superior performance across both numerical simulations and physical experiments. For the hardware tests, the researchers utilized a physical two-axis ISP mounted on a Stewart platform, a six-degree-of-freedom motion simulator that mimics the bumpy ride of a UGV.
- Dynamic Stiffness: The system's resistance to external forces increased significantly from 115 Nm/(rad/s) to 323 Nm/(rad/s).
- Tracking Accuracy: Frequency analysis showed a 55% improvement in relative angular rate tracking error.
- Real-World Performance: Experimental tests revealed a 28.3% improvement in angular position tracking and a 37.7% boost in overall stability accuracy.
- Transient Response: Settling time under step disturbances was reduced from 0.047s to 0.041s.
Crucially, using Nyquist plots, the authors showed that these modifications significantly improved performance while the stability margin remained almost identical to the standard version, proving that the gains did not come at the cost of system robustness.
Conclusion
The proposed modifications significantly enhance the performance of ISPs by closing the mathematical gap in disturbance rejection and accounting for sensor dynamics. While the current work focuses on first-order nonlinear ADRC for speed loops, the authors suggest that extending these correction techniques to higher-order systems is the next logical step for even more complex industrial applications.
Summary updated